JP6075042B2 - Language processing apparatus, language processing method, and program - Google Patents

Description

  The present invention relates to a language processing device, a language processing method, and a program.

  Regarding sentence reading support, a language processing method that displays the structure of a sentence in an easy-to-understand manner and helps understand the meaning is known. For example, when reading and understanding a language other than your native language, it is often difficult to understand the meaning of the word without knowing the structure of the sentence. Even in such a case, displaying the sentence structure in an easy-to-understand manner can help understanding.

  For example, referring to the analysis target classification knowledge from the information of the classification that is the subcategory of the phrase attached to the node to be analyzed in the parse tree, the analysis target is a verb phrase having the verb as the main word. There are known examples of determining whether a non-verb phrase having a noun as a main word or a part of speech. At this time, if the object to be analyzed is a verb phrase, the relationship between the verb and the deep case of the verb is analyzed. If the object to be analyzed is a non-verb phrase, a modification relationship including a noun phrase and an adjective phrase is obtained. To analyze. If the analysis target is a part of speech, the semantic analysis is terminated. In this way, analysis is performed by specifying the deep case necessary for the subsequent keyword extraction and generating the semantic structure of the minimum necessary range without performing other unnecessary semantic analysis inside the deep case. The goal is to save time. Another object of the present invention is to reduce the maintenance burden of the processing method by clarifying the flow of the entire analysis.

  Also, there is known an example in which search population data is generated from function model data in a knowledge classification support device that classifies knowledge about an object in accordance with the setting of search population setting data. This knowledge classification support device receives knowledge data related to a function node, searches for a function node similar to the knowledge data, and generates search result data in which information on the similar function node is stored. This is intended to facilitate the assignment of knowledge classifications to the function nodes constituting the function tree. When any information in the search result data is selected, this knowledge classification support device highlights a function node related to the arbitrary information on the display unit as a classification destination candidate node.

  Furthermore, a behavior data display method for displaying a graph that represents a behavior as a node and a transition of the behavior as an arc is known. In this behavior data display method, the behavior data display device calculates the initial probability of each behavior and the transition probability between the behaviors based on the behavior history of the behavior subject. Based on the calculated initial probability of each action and the transition probability between actions, the action data display device calculates the simplicity when the action or the transition from the action to another action is removed. Remove the smallest. The behavior data display device generates a graph in which the behavior is represented by a node and the transition between the behaviors is represented by an arc based on the simplified behavior and the transition probability between the behaviors. At this time, it is known that a plurality of actions having similar transition probabilities in the action history are grouped together as one action, and are grouped according to a predefined action hierarchy at the time of grouping. This aims to make it easier to find characteristic parts.

  Furthermore, in the virtual input device, when coordinates on the screen are input from the instruction unit to the control unit, and the button corresponding to the coordinates is the representative virtual button, the control unit selects the representative virtual button. Examples are known. In this example, the virtual input device transmits a command to set the virtual button belonging to the representative virtual button to the display state to the display state switching unit, and the display state switching unit sets the virtual button to the display state. When the button corresponding to the coordinates is a virtual button, the control unit checks the pressed state of the button attached to the instruction unit, and when the virtual button is pressed, the control unit performs the process assigned to the virtual button. Run. If there is no button corresponding to the coordinates, it is checked whether a representative virtual button in the selected state exists. If present, the control unit puts the representative virtual button into a non-selected state, and puts the virtual buttons belonging to the representative virtual button into a non-display state with respect to the display state switching unit. Accordingly, an object of the present invention is to provide a software keyboard that is easy to learn with a list property even with a small display area.

JP 2000-148755 A JP 2007-115139 A JP 2004-102681 A JP 2000-181608 A

  By the way, in the analysis of long sentences, it often takes time to understand which word is related to which word and which expression is related to which expression. However, the conventional semantic analysis and display methods as described above may be used to capture phrases globally, but cannot display the dependency on individual word levels. Therefore, it may not be possible to determine whether a meaningful word is a word necessary for decoding as a result of analysis, and display for appropriate decoding support may not be performed.

  Therefore, an object of the present invention is to provide a language processing apparatus, a language processing method, and a program that can easily display an analysis result even for a long sentence.

A language processing apparatus according to one aspect includes a phrase determination unit, an analysis unit, an expansion unit, a display processing unit, and a display unit. The phrase determining unit extracts a plurality of words included in the input sentence and analyzes the meaning of the sentence, and for each of the plurality of words, a phrase including the word includes a single meaning. It is determined whether or not there is, and if there is, the phrase is output. The analysis unit performs a morphological analysis of the sentence, performs a syntax analysis of the sentence based on a relationship between the morphemes of the sentence based on the morphological analysis, and based on a result of the syntax analysis, Generates relationship information indicating a semantic relationship between two related words and a semantic relationship between each of a plurality of the words and a word having a main meaning in the phrase output from the phrase determination unit. . The extension unit displays, as the relationship information and the separate phrase, whether or not to display the word or the phrase as another phrase connected to the preceding or following word or phrase based on the relationship information. This is performed according to extension information in which the relationship with availability is defined in advance . Display processing unit summarizes the word or phrase that is determined to be displayed as the specific phrase as a phrase. The display unit, based on the analysis result of analyzing the meaning of the sentence and the result of processing in the display processing unit, the words and phrases analyzed as the central concept of the sentence , the phrases summarized by the display processing unit, and The relationship information representing the semantic relationship between the phrase and the phrase is displayed.

Language processing method according to another aspect, language processing apparatus extracts a plurality of words included in the input sentence based on the results analysis to analyze the meaning of the sentence, for each of a plurality of said word, said It is determined whether or not there is a phrase that includes a word and has one meaning, and if there is, the phrase is output. The language processing apparatus performs morphological analysis of the sentence. Furthermore, the language processing apparatus performs syntax analysis of the sentence from the relationship between the morphemes of the sentence based on the morphological analysis. In addition, the language processing device , based on the result of the syntax analysis, in the semantic relationship between two words that are related to each other among the plurality of words, and each of the plurality of the words in the output phrase Relation information indicating a semantic relation with a word having a main meaning is generated. Further, the language processing device displays, as the relation information and the another phrase, whether or not to display the word or the phrase as another phrase connected to the preceding or following word or phrase based on the relation information. This is performed according to the extension information in which the relationship with whether or not to be allowed is defined in advance . Further, the language processing unit summarizes the word or phrase that is determined to be displayed as the specific phrase as a phrase. Further, the language processing unit, based on the results of collectively Ru processing as the analysis result and the one phrase meaning analyzed the sentence, phrase parsed as central idea of the sentence, the gathered phrase and, The relationship information representing the semantic relationship between the phrase and the phrase is displayed.

  In addition, even if it is a program for making a computer perform the method which concerns on the aspect mentioned above, since there exists an effect similar to the method which concerns on the aspect mentioned above by making this program run by the said computer, the subject mentioned above Is resolved.

  According to the above-described aspect, it is possible to provide a language processing device, a language processing method, and a program that can easily display an analysis result even for a long sentence.

It is a block diagram which shows an example of the function of a language processing apparatus. It is a figure which shows an example of graph expression. It is a figure which shows an example of the reading assistance in a graph expression. It is a figure which shows another example of graph expression. It is a figure which shows degeneration of a node. It is a figure which shows the example of a display of an input sentence. It is a figure which shows an example of morphological analysis information. It is a figure which shows an example of part of speech symbol information. It is a figure which shows an example of parsing information. It is a figure which shows an example of semantic analysis information. It is a figure which shows an example of extended information. It is a figure which shows an example of the initial display of an analysis result. It is a figure which shows another example of the initial display of an analysis result. It is a figure which shows the example of expansion | deployment of a graph. It is a figure which shows another example of expansion | deployment of a graph. It is a figure which shows another example of expansion | deployment of a graph. It is a figure which shows the process which calculates a response | compatibility with an input sentence character string and a node. It is a figure which shows an example of the determination process of a noun phrase and a verb phrase. It is a figure which shows determination of the magnitude | size of the noun phrase containing a specific character string. It is a figure which shows the example of a display at the time of expanding a node to a noun phrase. It is a figure which shows the example of a display when a user selects a specific character string in an input sentence. It is a figure which shows another example of a display when a user selects a specific character string in an input sentence. It is a figure which shows an example of expansion of a node. It is a figure which shows an example of expansion of a node. It is a figure which shows the example which simplified the example of a display of graph expression by expansion of the node. It is a figure which shows an example of graph expression. It is a main flowchart which shows operation | movement of a language processing apparatus. It is a flowchart which shows a node expansion | deployment and degeneracy process. It is a flowchart which shows a node expansion process. It is a flowchart which shows a node expansion determination process. It is a figure which shows the modification of a display method. It is a figure which shows the modification of a display method. It is a figure which shows the example of a display at the time of the character string selection. It is a figure which shows the example of a display at the time of the character string selection. It is a figure which shows the hardware constitutions of a standard computer.

(First embodiment)
The language processing apparatus 1 according to the first embodiment will be described below with reference to the drawings. FIG. 1 is a block diagram illustrating an example of functions of the language processing device 1. The language processing device 1 is a device that analyzes an input sentence and displays the meaning and structure of the sentence in a simple expression.

  As shown in FIG. 1, the language processing device 1 includes an input unit 3, an analysis unit 5, a noun phrase determination unit 13, a verb phrase determination unit 15, a node expansion unit 17, a phrase display optimization unit 19, a correspondence calculation unit 21, A display control unit 23, an analysis result storage unit 25, and a display unit 27 are provided. The analysis unit 5 includes a morphological analysis unit 7, a syntax analysis unit 9, and a semantic analysis unit 11.

  The input unit 3 is an input device such as a keyboard or a touch panel. The analysis unit 5 analyzes the input sentence and outputs an analysis result. In the analysis unit 5, the morpheme analysis unit 7 performs morpheme analysis of the input sentence and outputs morpheme analysis information 150. The syntax analysis unit 9 performs syntax analysis based on the morphological analysis information 150 and outputs syntax analysis information 180. The semantic analysis unit 11 performs semantic analysis based on the morphological analysis information 150 and the syntax analysis information 180 and outputs the semantic analysis information 200.

  The noun phrase determination unit 13 determines words constituting the noun phrase based on the syntax analysis information 180. The verb phrase determination unit 15 determines the phrases constituting the verb phrase based on the syntax analysis information 180. The node expansion unit 17 determines words to be displayed together based on the determination results of the noun phrase determination unit 13 and the verb phrase determination unit 15. The phrase display optimizing unit 19 performs processing for grouping phrases determined to be displayed together by the node expanding unit 17 into one phrase. The correspondence calculation unit 21 is used for graph expression of an input sentence, which will be described later, in accordance with outputs from the input unit 3, noun phrase determination unit 13, verb phrase determination unit 15, node expansion unit 17, and phrase display optimization unit 19. Calculation processing, calculation processing for associating a character string of an input sentence with a node in a graph expression, and the like are performed.

  The display control unit 23 controls the display unit 27 based on outputs from the analysis unit 5, noun phrase determination unit 13, verb phrase determination unit 15, node expansion unit 17, phrase display optimization unit 19, and the like. The display unit 27 is a display device such as a liquid crystal display device, for example. The analysis result storage unit 25 is a storage device that stores morphological analysis information 150, syntax analysis information 180, semantic analysis information 200, and the like.

  A graph representation of the relationship between words used in the present embodiment will be described with reference to FIGS. FIG. 2 is a diagram illustrating an example of a graph expression. As shown in FIG. 2, the expression example 30 is an example in which a sentence “Taro goes to school tomorrow” is expressed in a graph. Here, the graph expression indicates the concept of words constituting a sentence by “nodes”, and the relationship between words is expressed by “arc” indicating the direction of the arrow and the content of the relationship. The concept represented by the node is originally represented by a semantic symbol corresponding to the word. However, in the present invention, for the purpose of easily displaying it to the user, it is not a concept but a word (or a phrase or phrase having the word as the main word) itself. It shall be indicated by

  The sentence “Taro will go to school tomorrow” has the concept of “Taro”, “Tomorrow”, “School”, and “Row”, and corresponding nodes 32 to 38 are set. For example, in the expression example 30, when attention is paid to the “row” of the node 32, the relationship between words is represented by an arc 44 of “target” toward the “school” of the node 38. The arc 44 indicates that “school” represents the goal (destination) of “line”. An arc 42 “time” appearing from “row” of node 32 to “tomorrow” of node 36 indicates that “tomorrow” represents the time of “row”. Similarly, the “actor” of the arc 40 appearing from the “row” of the node 32 to the “taro” of the node 34 represents that “taro” is the subject of the “row” operation. Hereinafter, the concept represented by the node is referred to as the node content, and the relationship between the words indicated by the arc is referred to as the arc content.

  FIG. 3 is a diagram illustrating an example of reading support in the graph expression. As shown in FIG. 3, for example, in the sentence 52, when “school” of the selection unit 54 is selected by the user via the input unit 3 as indicated by an arrow 56, the node 38 is like the selection node 37. For example, reading is supported by highlighting. At this time, the selection may be performed when the node 38 of the expression example 30 is selected like the selection 58 by a touch panel or the like.

  FIG. 4 is a diagram illustrating another example of the graph expression. Expression example 60 is an example in which the structure of “Taro goes to school where Hanako goes tomorrow” is expressed in a graph. Thus, when there are modifiers, the graph becomes complicated. In the expression example 60, nodes 32 to 34, nodes 62 and 64 having contents of “row”, “taro”, “tomorrow”, “school”, “row”, and “Hanako”, and arcs 40 to 44, 66 indicating their relationship, 68 is shown.

  However, if all the nodes are expanded and displayed as in the expression example 60, the expression becomes complicated and the purpose of supporting reading comprehension cannot be achieved. For this reason, a node with a lot of information can be degenerated. Note that expanding a node means displaying a plurality of nodes connected to the node separately, and degenerating a node means that one node and at least one node connected to the node are It means to display on behalf of one node. At this time, the node displayed as a representative is called a degenerate node.

  FIG. 5 is a diagram illustrating node degeneration. The expression example 70 is an example in which the structure representing “the school where Hanako goes” is reduced to “school”. “School” is, for example, displayed by a reduced node 39 in which hatching, color, or the like is changed, and represents a reduced node. Here, if the user selects the reduced node 39 having “school” as a content by, for example, a touch panel or a click operation, the reduced node 39 of “school” is expanded, for example, node 62 (“row”) is expanded. Displayed. In this way, if a user selects and expands a node that he / she wants to read deeply, reading support can be provided.

  Next, the language processing method according to the first embodiment will be further described with reference to FIGS. 6 to 29, taking another input sentence as an example. FIG. 6 is a diagram illustrating a display example of the input sentence 80. The input sentence 80 is a slightly more complicated sentence than the example described with reference to FIGS.

  As shown in FIG. 6, the input sentence 80 is the following sentence. “This technology uses a co-occurrence relationship in which each element that delimits the document before proofreading in a predetermined unit appears in the document before proofreading, and each element that delimits the document after proofreading in a predetermined unit in the document after proofreading. Based on the co-occurrence storage unit that stores the co-occurrence relationship that is the relationship that appears together, the frequency of occurrence of the co-occurrence relationship of each element in the document before proofreading and the frequency of occurrence of the co-occurrence relationship of each element in the document after proofreading, respectively Hereinafter, a case where the input sentence 80 is analyzed will be described.

  FIG. 7A is a diagram illustrating an example of the morphological analysis information 150. The morpheme analysis is a process of analyzing morphemes constituting a sentence with reference to a dictionary or the like provided in advance in the analysis result storage unit 25 or the like. The morpheme analysis information 150 is an example showing a part of the morpheme analysis result of the input sentence 80, and is generated by the morpheme analysis unit 7, for example. The morpheme analysis information 150 indicates a morpheme number 152, a morpheme 154, a start position 156, a character length 158, part of speech information 160, and a semantic symbol 162. The morpheme number 152 is a morpheme identification number, and the morpheme 154 is a morpheme character string. The start position 156 is the number of characters up to the start position of the morpheme in the input sentence, and the character length 158 is the number of characters of the morpheme character string. The part of speech information 160 is a symbol indicating the part of speech of the morpheme, and the semantic symbol 162 is a symbol indicating the meaning of the morpheme.

  For example, the morpheme number 152 = “1” is the morpheme 154 = “book”. This morpheme start position 156 = "0", indicating that the head of the input sentence 80 is the start position. Character length 158 = “1”, part-of-speech information 160 = “DET”, and semantic symbol 162 = “THIS”.

  FIG. 7B is a diagram illustrating an example of the part of speech symbol information 170. The part-of-speech symbol information 170 is information indicating a symbol representing a part-of-speech morpheme, and is stored in advance in the analysis result storage unit 25, for example. As shown in FIG. 7B, for example, the part of speech symbol 172 = “DET” represents an article, “N” represents a noun, and “J” represents a particle. Therefore, “book” in FIG. 7A is determined to be “article”.

  FIG. 8 is a diagram illustrating an example of the syntax analysis information 180. The syntax analysis is a process of analyzing the relationship between morphemes constituting a sentence into meaningful phrases with reference to a dictionary or the like provided in advance in the analysis result storage unit 25 or the like. As shown in FIG. 8, the syntax analysis information 180 is information indicating an example of the result of syntax analysis, and is generated by the syntax analysis unit 9. The parsing information 180 includes a component 182, a part of speech 184, a label 186, and a main word 188. The component 182 indicates the morpheme number 152 analyzed by the morpheme analysis information 150 and the label 186. The part of speech 184 indicates the part of speech of the component 182. The label 186 is a name given to the component 182. The main word 188 is a morpheme number 152 indicating a word holding the main meaning in the component 182.

  FIG. 9 is a diagram illustrating an example of the semantic analysis information 200. Semantic analysis is a process of analyzing the relationship between semantic symbols of words from a syntax analysis result with reference to a dictionary or the like provided in advance in the analysis result storage unit 25 or the like. As shown in FIG. 9, the semantic analysis information 200 is information indicating an example of the result of semantic analysis, and is generated by the semantic analysis unit 11. The semantic analysis information 200 includes a departure node 202, a reaching node 204, and an arc 206. The departure node 202 shows a morpheme number 208, a semantic symbol 210, and a morpheme 212, and the destination node 204 shows a morpheme number 214, a semantic symbol 216, and a morpheme 218. The departure node 202 is a node where the arc is generated, and the reaching node 204 is a node where the arc is directed. The morpheme number 208 is an identification number of the morpheme corresponding to the contents of the node. The meaning symbol 210 is a symbol indicating the meaning of the morpheme. The morpheme 212 is a morpheme corresponding to the contents of the departure node 202. The same applies to the morpheme number 214, the semantic symbol 216, and the morpheme 218 of the destination node 204.

  The semantic analysis information 200 is an example in which a character string “based on a stored co-occurrence storage unit” that is a part of the input sentence 80 is analyzed. In the semantic analysis information 200, for example, the morpheme number 208 = “65” of the departure node 202 is the semantic symbol 210 = “BASED” and the morpheme 212 = “based”. On the other hand, the morpheme number 214 = “64” of the reaching node 204 is the semantic symbol 216 = “PART”, and the morpheme 218 = “part”. At this time, the arc 206 = “modification”, and “part” indicates that “based” is “modified”. Based on the analysis result, for example, a graph is created in which an arc having “modification” as a content is directed from a node having “based” as content to a node having “part” as content. Similarly, a graph indicating the input sentence 80 is created for other words.

  FIG. 10 is a diagram illustrating an example of the extended information 230. The extension information 230 is information indicating the relationship between the contents of the arc and whether or not two nodes connected by the arc are grouped together (hereinafter, grouping the nodes together is referred to as node expansion). For example, it is stored in advance in the analysis result storage unit 25.

  The extended information 230 has an arc 232 and a type 234. The arc 232 indicates the content of the arc, and the type 234 indicates the type of whether or not nodes connected to the arc can be expanded. As illustrated in FIG. 10, for example, the arc 232 (“modification”) has the type 234 = “expansion”, and indicates that two nodes in the “modification” relationship can be expanded.

  Next, display examples of the analysis result of the input sentence 80 will be described with reference to FIGS. FIG. 11 is a diagram illustrating an example of an initial display of analysis results. As shown in FIG. 11, in this example, a display example 82 is shown. The display example 82 includes a display example 86 and a graph 83. The graph 83 is a graph representation of the input sentence 80, and one degenerate node 84 is displayed. The display example 86 is a sentence display of the input sentence 80, and “Calculate” is highlighted 88. “Calculate” is a phrase that is analyzed as a central concept of the input sentence 80 by analyzing the input sentence 80 by the analysis unit 5.

  FIG. 12 is a diagram illustrating another example of the initial display of the analysis result. As shown in FIG. 12, a display example 90 is shown in this example. The display example 90 includes a display example 91 and a graph 92. A graph 92 is a graph representation of the input sentence 80 and shows an example in which the degenerate node 84 is expanded. In the graph 92, the degenerated node 84 is expanded and changed to the display of the node 85. A degeneration node 92 is connected to the node 85 by an arc 100, a degeneration node 94 is connected by an arc 102, a degeneration node 96 is connected by an arc 104, and a degeneration node 98 is connected by an arc 106.

  In the graph 92, the contents of the degeneration node 92 are “technology”, the contents of the degeneration node 94 are “based”, the contents of the degeneration node 98 are “appearance frequency”, and the contents of the degeneration node 98 are “respectively”. It is. The content of the arc 100 is “operator”, the content of the arc 102 is “modification”, the content of the arc 104 is “target”, and the content of the arc 106 is “method”. The contents of the nodes and arcs in the graph 92 and the contents of other nodes and arcs to be described later are analyzed by the morphological analysis information 150, the syntax analysis information 180, and the semantic analysis information 200.

  The display example 91 is a sentence expression. In addition to the highlight display 88, the words “this technology” is displayed as the highlight display 93, “based on” as the highlight display 95, and “appearance frequency” are displayed as the highlight display 97, respectively. Has been. These highlights 93 to 97 correspond to the nodes 92 to 98.

  In the graph 92, when a node with an arc such as the node 85 is selected, the node is degenerated and the display returns to the degenerated node display. In this example, the node 85 returns to the display of the degenerated node 84, and the connected arcs 100 to 106 and nodes 92 to 98 are not displayed. On the other hand, when a degenerate node is selected, for example, an arc directly coming out from the degenerate node and a node connected thereto are displayed.

  FIG. 13 is a diagram illustrating a development example of the graph. As illustrated in FIG. 13, the graph 110 is a state in which the degenerate node 96 (“appearance frequency”) is selected in the graph 92. In the graph 110, the degenerate node 96 is expanded to display the node 97, and the arc 120 directly appearing, the degenerate node 114 connected to the arc 120, the arc 118, and the node 112 connected to the arc 120 are shown. In graph 110, arc 122 and degenerate node 116 exiting node 112 are shown. In this way, the degenerate node 96 is expanded as in the expansion example 126, for example. At this time, although not shown in the graph 110, in the input sentence 80, it is preferable to highlight a portion corresponding to the expanded example 126, such as the display example 86 of FIG.

  In the graph 110, the content of the node 112 is “appearance frequency”, and the content of the degenerate nodes 114 and 116 is “co-occurrence relationship”. The contents of the arc 118 are “enumeration”, and the contents of the arcs 120 and 122 are “modification”.

  FIG. 14 is a diagram illustrating another development example of the graph. As illustrated in FIG. 14, the graph 130 is a state in which the degenerate node 94 (“based”) is selected in the graph 110. In the graph 130, the degenerated node 94 is expanded to become a display of the node 95, and the arc 134 directly appearing and the degenerated node 132 connected thereto are shown. In this way, the degenerate node 94 is expanded as in the expansion example 136, for example. In the graph 130, the content of the node 132 is “part”, and the content of the arc 134 is “modification”.

  At this time, the expanded example 126 expanded in the graph 110 of FIG. 13 has returned to the reduced node 96 as in the reduced example 138. As a result, only the part to be noticed is displayed in detail, and the display is simpler.

  FIG. 15 is a diagram illustrating another development example of the graph. A graph 220 in FIG. 15 is an example of a graph expression in which the graph 130 in FIG. 14 is further developed. As illustrated in FIG. 15, the graph 220 is a graph representation in which the degenerated node 132 (“part”) is expanded in the graph 130.

  In the graph 220, the degenerated node 132 is expanded to display a node 133, and the arc 225 directly appearing and the degenerated node 221 connected thereto are shown. Also shown are an arc 226 and a degenerate node 222 connected thereto, an arc 227 exiting therefrom and a node 223 connected thereto. In this way, the degenerate node 94 is expanded as in the expansion example 228, for example.

  In the graph 220, the content of the node 222 is “memory”, the content of the node 223 is “co-occurrence”, and the content of the degenerate node 221 is “storage”. The content of the arc 225 is “operator”, the content of the arc 226 is “modification”, and the content of the arc 227 is “target”.

  Subsequently, the language processing according to the first embodiment will be further described with reference to FIGS. 16 to 23. FIG. 16 is a diagram showing processing for calculating the correspondence between the input sentence character string and the node. In FIG. 16, a case where a sentence 240 "This is a pen" is given as an input sentence will be described. The sentence 240 is analyzed by the analysis unit 5 into morphemes 242 to 250 of “this”, “ha”, “pen”, “is”, and “.”.

  Among these morphemes 242-250, for example, “this” of the morpheme 242 and “pen” of the morpheme 246 are associated with the nodes 243 and 247, and the relationship between these nodes is “<ISA>”. It is represented by At this time, the morpheme 242 has a start position (offset) of “0” and the number of characters of “2”. The morpheme 246 has a start position (offset) of “3” and the number of characters of “2”. In this way, the correspondence calculation unit 21 calculates the correspondence between each node of the analysis result of the input sentence and the character string received by the input unit 3.

  FIG. 17 is a diagram illustrating an example of determination processing for a noun phrase and a verb phrase. As shown in FIG. 17, in the phrase determination example 250, the noun phrase determination unit 13 and the verb phrase determination unit 15 perform the determination process on the character string 252 “in the co-occurrence storage unit storing“ ˜ ”and“ ... ”. . The part of speech for each morpheme in the character string 252 is represented by the part of speech information 254. This part-of-speech information 254 is obtained from the analysis result by the morphological analysis unit 7.

  For example, a character string “co-occurrence storage unit” in a part of the character string 252 is represented by three nouns (N) “co-occurrence”, “memory”, and “part” as shown in the part-of-speech information 254. At this time, as shown in the phrase information 256, the noun phrase determination unit 13 determines that “co-occurrence” and “memory” constitute one noun phrase (NP) “co-occurrence memory”. Furthermore, it is determined that “co-occurrence memory” and “part” constitute one noun phrase. A minimum noun phrase 270 illustrated in FIG. 17 is a minimum noun phrase including “part”.

  Similarly, as indicated by the phrase information 256, “to and” and “... And” are both noun phrases. Furthermore, as shown in the phrase information 258, “to,...,” Is determined to be a noun phrase. As shown in the phrase information 260, “to and...” Are also determined to be noun phrases.

  “Stored” is a verb (V) as shown in the part of speech information 254. “Stored with ˜ and...” Is determined as a verb phrase (VP) by the verb phrase determination unit 15 as shown in the phrase information 262.

  Furthermore, as shown in the phrase information 264, the “co-occurrence storage unit storing“ ˜ ”and“... ”Is determined to be a noun phrase. Similarly, as shown in the phrase information 266, the character string 252 “in the co-occurrence storage unit storing ˜ and...” Is determined to be a noun phrase.

  FIG. 18 is a diagram illustrating determination of the size of a noun phrase including a specific character string. Phrase determination example 280 shows a determination example of the size of a noun phrase including “part”. The “part” is the “main word” of the character string 252 “in the co-occurrence storage unit storing“ ˜ and.

  As shown in FIG. 18, a noun phrase including “part” refers to phrase information of a character string including “part” from the bottom to the top of FIG. "Is a character string corresponding to the place where" According to the arrow 282, the minimum character string including “part” is the minimum noun phrase 270 (“co-occurrence storage unit”). Similarly, the second smallest noun phrase refers to phrase information as shown by an arrow 284 in the upward direction in FIG. Thus, the second smallest noun phrase is the noun phrase 272 (“co-occurrence storage unit storing“ ˜ ”and“... ”).

  FIG. 19 is a diagram illustrating a display example when a node is expanded to a noun phrase. As illustrated in FIG. 19, the display example 290 includes a display example 292 and a graph 296. The display example 292 is an example in which the selected character string 294 selected by the user in the input sentence 80 is highlighted with an underline.

  A graph 296 is an example showing a graph expression of the analysis result corresponding to the selected character string 294 selected. For example, in the graph 130 of FIG. 14, a simplified node 298 (“co-occurrence storage unit”) corresponding to the minimum noun phrase including “part” is shown instead of “part” represented by the degenerate node 132. Yes. A simplified node refers to a node that is displayed corresponding to an expanded phrase.

  20 and 21 are diagrams showing another display example when the user selects a specific character string in the input sentence 80. As illustrated in FIG. 20, in the display example 310, it is assumed that the character string “frequency and proofread” is selected by the user. The selected selected character string 312 is highlighted, for example.

  By the way, let us consider making this character string “frequency and proofread” into a graph expression. Since “frequency and proofreading” are directly connected in the input sentence 80, they can be selected as in the display example 310. However, in the graph representation, “frequency” and “calibration” are not directly connected. In addition, “frequency” is a word in the noun phrase “appearance frequency”, and “proofreading” is a word in the noun phrase “document after proofreading”. In such a case, the selected portion is automatically expanded.

  FIG. 21 is a diagram illustrating a display example of analysis results. In FIG. 21, a display example 320 includes a display example 322 and a graph 340. The display example 322 is an example in which, in the input sentence 80, a character string to be noticed when reading support for the selected character string 312 selected by the user in FIG. A graph 340 is a diagram illustrating an example of a graph expression in which the graph 110 in FIG. 13 is further developed.

  In the graph 340, the degenerate node 116 in the graph 110 is expanded to become a node 117, and an arc 350, a node 344 connected thereto, an arc 352, and a node 348 connected thereto are displayed. The content of the node 344 is “document after proofreading”, and the content of the node 348 is “each element”. The content of the arc 350 is “range”, and the content of the arc 352 is “related”. In this graph 340, expansion targets 354, 356, and 358 are set.

  22 and 23 are diagrams illustrating an example of node expansion. FIG. 22 illustrates, for example, node expansion in the expansion target 356 illustrated in FIG. As shown in FIG. 22, in the expansion target 356, the “appearance frequency” of the node 112 and the “co-occurrence relationship” of the degenerate node 116 are displayed separately. To do. The content of the simplified node 372 is “appearance frequency of co-occurrence relationship”.

  FIG. 23 illustrates node expansion in the expansion target 358. As shown in FIG. 23, in the expansion target 358, the “post-proof document” of the node 344 and the “each element” of the node 348 are separately displayed, and these are collectively referred to as one simple node 377. The contents of the simplified node 377 are “each element in the proofread document”.

  FIG. 24 is a diagram illustrating an example in which the graph representation graph 340 in FIG. 21 is simplified by node expansion. As illustrated in FIG. 24, the display example 380 includes a display example 382 and a graph 390. The display example 382 is an example in which, in the input sentence 80, a portion that assists reading comprehension of “frequency and proofreading” selected by the user is highlighted. In the display example 382, highlighted displays 384, 328, 386, and 388 are shown.

  In the graph 390, extended nodes 372, 379, and 392 are displayed. By these displays, noun phrases are displayed together, the display location may be narrow, and the display is simple.

  FIG. 25 is a diagram illustrating an example of a graph expression. As shown in FIG. 25, the display example 490 includes a character string 492 and a graph 500. The character string 492 is “appearance frequency of the co-occurrence relationship of each element in the proofread document”, and the graph 500 shows the analysis result of the character string 492 in a graph expression. In the graph 500, the node 344 in FIG. 23 is expanded into a node 502 (“element”) and a node 504 (“each”), and a node 348 includes a node 494 (“document”) and a node 496 (“after”). , Node 498 (“calibration”).

  Next, the operation of the language processing apparatus 1 according to the first embodiment will be described with reference to FIGS. 26 is a main flowchart showing the operation of the language processing apparatus, FIG. 27 is a flowchart showing node expansion and reduction processing, FIG. 28 is a flowchart showing node extension processing, and FIG. 29 is a flowchart showing node extension determination. . The operation of the language processing apparatus 1 will be described on the assumption that each function described with reference to FIG. 1 performs processing.

  First, the overall flow of processing of the language processing device 1 will be described with reference to FIG. As illustrated in FIG. 26, the analysis unit 5 detects that an input sentence has been input via the input unit 3 by the user, for example. Alternatively, the analysis unit 5 reads an input sentence to be analyzed from, for example, a storage device including the analysis result storage unit 25. The analysis unit 5 analyzes the input sentence that has been input (S401). At this time, the morpheme analysis unit 7 performs morpheme analysis and generates, for example, morpheme analysis information 150 (S411). The syntax analysis unit 9 performs syntax analysis based on the morphological analysis information 150, generates syntax analysis information 180 (S412), and stores it in the analysis result storage unit 25 (S413). At this time, the analysis unit 5 may store the morphological analysis information 150 in the analysis result storage unit 25.

  The semantic analysis unit 11 performs semantic analysis based on the morphological analysis information 150 and the syntax analysis information 180, generates semantic analysis information 200 (S414), and stores it in the analysis result storage unit 25 (S415).

  The correspondence calculation unit 21 generates information for graph expression based on the semantic analysis information 200 and calculates the correspondence between the node and the character string. The display control unit 23 causes the display unit 27 to perform initial display of the analysis result based on the calculation result in the correspondence calculation unit 21 (S402). The initial display is, for example, the display example 82 in FIG. 11, the display example 90 in FIG. 12, and the like, and is an input sentence partially highlighted or a graph representation display. The correspondence between the node and the character string is calculated based on the analysis result of the morpheme analysis unit 7 or the like as described with reference to FIG.

  The correspondence calculation unit 21 determines whether it is detected that the user has selected the end of the process via the input unit 3 (S403), and when it is detected that the user has selected (S403: YES), the process ends. To do.

  When it is not detected that the end is selected (S403: NO), the correspondence calculation unit 21 determines whether a character string or a node is selected in the display of the input sentence such as the display example 82 and the display example 90, for example. Repeat (S404: NO). At this time, the selected node may be highlighted so as to indicate that it has been selected. When the correspondence calculation unit 21 determines that a character string or a node is detected (S404: YES), the node expansion unit 17 performs node expansion and contraction processing shown in FIG. 27 (S405).

  At this time, for example, when a character string is selected, the correspondence calculation unit 21 specifies the selected node by calculating the correspondence between the character string and the node. For example, when the character string “frequency and proof” is selected as in the example of FIG. 20, the correspondence calculation unit 21 corresponds to “frequency of appearance” of the degenerate node 96, and “proof” is the node It is specified that it corresponds to “Post-proofreading document” 344.

  Next, the node expansion / reduction process will be described with reference to FIG. As shown in FIG. 27, the node expansion unit 17 acquires a set of nodes directly connected from the selected node, and attaches a label to each (S421). The labels can be, for example, expansion target nodes N1 to Nn (n is an integer of 1 or more). When expanding the node, the node expansion unit 17 performs the node expansion process shown in FIG. 28 for all the expansion target nodes N1 to Nn acquired in S421 (S422).

  Hereinafter, a case where “based” of the degenerate node 94 in the display example 90 of FIG. 12 is selected in S404 of FIG. 26 will be described as an example. In the example of FIG. 12, all nodes that have been reduced to the selected reduced node 94 are acquired. For example, as shown in FIG. 15, nodes 133, 221, 222, and 223 are degenerated in the degeneration node 94. However, only the node 133 is directly connected by an arc. Therefore, the node 133 is set as the expansion target node N1. That is, the set of nodes to be acquired is the node 133 (“part”).

  Returning to FIG. 27, the phrase display optimizing unit 19 performs a process of collecting noun phrases and verb phrases expanded by the node expansion process (S423). That is, as described with reference to FIG. 18, when the noun phrase determining unit 13 extracts the minimum noun phrase “co-occurrence storage unit” for the expansion target node N1 = node 133 (“part”). The phrase display optimization unit 19 performs a process in which the “co-occurrence storage unit” is one node. At this time, the phrase display optimizing unit 19 determines how far the node is expanded based on the expansion processing by the node expansion unit 17. The display control unit 23 displays the nodes of the “co-occurrence storage unit” expanded based on the processing results of the phrase display optimization unit 19 and the correspondence calculation unit 21, for example, a display indicating that a highlight display or the like has been selected. (S424).

  Next, the node expansion process will be described with reference to FIG. As illustrated in FIG. 28, the node expansion unit 17 sets a processing target node N in the node expansion processing (S431). As described above, the processing target node N is the expansion target node N1 = node 133 (“part”).

  The noun phrase determination unit 13 and the verb phrase determination unit 15 determine whether or not the processing target node N is included in the noun phrase or the verb phrase (S432). When the target processing node N is included in the noun phrase or verb phrase (S432: YES), the noun phrase determination unit 13 or the verb phrase determination unit 15 determines the main word of the noun phrase or verb phrase including the processing target node N. A node (hereinafter referred to as the main node NA) is automatically selected (S433). When the target processing node N is not included in the noun phrase or the verb phrase (S432: NO), the node expansion unit 17 advances the process to S423 in FIG.

  For example, as described with reference to FIG. 18, it is determined that “part” is included in the noun phrase “co-occurrence storage unit”. This determination is acquired by referring to the syntax analysis information 180, for example. The node extension unit 17 selects the main node NA = “part” of the noun phrase “co-occurrence storage unit”.

  Next, the node expansion unit 17 enumerates arcs output from the selected main word node NA and assigns identification numbers (ID) = 1, 2,..., K (k is an integer of 1 or more), respectively. The counter value ac = 0 is set (S434).

  As shown in FIG. 15, arcs exiting from the main node NA = “part” are an arc 225 (“operation main”) and an arc 226 (“modification”). Therefore, for example, ID = 1 is assigned to the arc 225 (“operator”), and ID = 2 is assigned to the arc 226 (“modification”).

  Returning to FIG. 28, the node expansion unit 17 expands the processing target node to the minimum noun phrase including the processing target node N and the main word node NA (S435). For example, when the processing target node N = “part”, the minimum noun phrase is “co-occurrence storage unit”.

  Subsequently, the node expansion unit 17 performs the node expansion determination illustrated in FIG. 29 (S436). As a result of this determination, if it is determined that further expansion is to be performed (S437: YES), the node expansion / reduction process shown in FIG. 27 is further performed on the node connected to the arc output in the process of FIG. S438). If it is determined in S437 that no further expansion is to be performed (S437: NO), the node expansion unit 17 advances the processing to S423 in FIG.

  Further, the node extension determination process will be described with reference to FIG. As illustrated in FIG. 29, the node expansion unit 17 determines whether or not the processing has been completed for all the arcs listed in S434 (S441). When the processing is completed (S441: YES), the node expansion unit 17 returns “NO” to the determination of S437 in FIG. 28 (S447).

  If the processing has not ended (S441: NO), the node expansion unit 17 sets the counter value ac = ac + 1 (S442), and determines whether or not the arc with ID = ac is the processing target arc (S443). . Here, the processing target arc is an arc that is not included in the minimum noun phrase determined in the node expansion processing of FIG.

  For example, when the minimum noun phrase is “co-occurrence storage unit”, arc 225 (“actor”) that is an arc with ID = 1 is not included in “co-occurrence storage unit”. is there. The arc with ID = 2 is the arc 226 (“modification”), which connects “part” and “memory”, and is included in the “co-occurrence storage unit”. Therefore, it is not a processing target arc.

  When the arc with ID = ac is not the processing target arc (S443: NO), the node expansion unit 17 returns “YES” in response to the determination of S437 of FIG. 28 (S446). When the arc is a processing target arc (S443: YES), the node expansion unit 17 determines whether or not the type of the target arc (ID = ac) is “extended” (S444). When the type of the target arc is not “expansion” (S444: NO), the node expansion unit 17 advances the process to S446. When the type of the target arc is “expansion” (S444: YES), the node expansion unit 17 outputs the target arc, proceeds to S446, and returns “YES” to the process of S437 in FIG. .

  For example, when the target arc is the arc 225 (“operator”) whose ID = 1 in FIG. 28, referring to the extended information 230, the “operator” is an “non-extended” arc. The node extension unit 17 proceeds with the process to S446 as it is. In the process of S437 of FIG. 28, when “YES” is acquired, that is, when the type of the target arc is “expansion”, the node expansion unit 17 outputs the target arc (S445). The node expansion unit 17 performs the node expansion / reduction process shown in FIG. 27 again for the node directly connected to the output target arc (S438). In this way, when the process is completed for all the arcs of the IDs, the node expansion unit 17 returns the process to S403 in FIG. Through the above processing, for example, as shown in FIG. 19, instead of the degenerate node 132 (“part”), the simplified node 298 (“co-occurrence storage unit”) is displayed, for example, highlighted.

  Similar processing will be described with reference to the example shown in FIG. In the example of FIG. 20, the input sentence 80 is analyzed (S401, S402), and the selected character string 312 = “frequency and proofreading” is selected as shown in the display example 310 (S403: NO, S404: YES). That is, when the selected character string 312 is selected via the input unit 3, the correspondence calculating unit 21 calculates the relationship between the selected character string and the node, thereby specifying the node from the selected character string. At the same time, the display control unit 23 causes the display unit 27 to display.

  As described above, “frequency” is a word in the noun phrase “appearance frequency”. Also, there is no node corresponding to “and”. Therefore, in such a case, the nodes selected in S404 in FIG. 26 are the degenerate node 96 (“appearance frequency”) (see FIG. 21) and the node 498 (“calibration”) (see FIG. 25). .

  Subsequently, the node expansion unit 17 acquires a set of nodes directly connected to the degenerate node 96 (“appearance frequency”) and the node 498 (“proofreading”) (S421). First, the degenerate node 96 (“appearance frequency”) will be described.

  As shown in FIG. 21, nodes directly connected from the degenerate node 96 are a degenerate node 114 (“co-occurrence relationship”) and a node 112 (“appearance frequency”). The node expansion process of FIG. 28 is performed with these nodes as the expansion target nodes N1 and N2 (S422).

  First, the node expansion unit 17 sets processing target node = expansion target node N1 = degenerated node 114 (“co-occurrence relationship”) (S431). The “co-occurrence relationship” is determined to be included in the noun phrase “appearance frequency of the co-occurrence relationship” by the noun phrase determination unit 13 (S432: YES). The main word of this noun phrase is determined to be the node 97 (“appearance frequency”) (S433). At this time, the arcs directly coming out from the main node are the arc 120 (“modification”) and the arc 118 (“enumeration”) (S434). Therefore, IDs are assigned to these arcs. The degenerate node 114 is expanded to the minimum noun phrase “occurrence frequency of co-occurrence relationship” including this main word (S435).

  29, since the “co-occurrence relationship” connected to the arc 120 (for example, ID = 1) is included in the expanded noun phrase “occurrence frequency of the co-occurrence relationship”, the processing target of the expansion (S441 to S443: NO). Therefore, the expansion related to the arc 120 is completed at the simplified node 392 (“occurrence frequency of co-occurrence relationship”) shown in FIG.

  Next, the process of FIG. 29 is performed for the arc 118 (“enumeration”) (for example, ID = 2). According to FIG. 10, “enumeration” is an arc of “expansion” (S444: YES). Therefore, the node expansion unit 17 outputs the arc “enumeration” (S445), and returns “YES” to the S437 process of FIG. 28 (S446). The node expansion unit 17 performs the process of FIG. 27 for the node 112 (“appearance frequency”) connected to the arc 118 (S438).

  The node expansion unit 17 acquires the node 112 as a node directly connected to the selected node (S421). The node expansion unit 17 performs the processing in FIG. 28 with the node 112 as a processing target node (S422, S431).

  The node 112 (“appearance frequency”) is included in the noun phrase “appearance frequency of co-occurrence relations” (S432: YES). The main word is the node 112 (“appearance frequency”) (S433). The arc 122 exiting from the node 112 is set to ID = 1 (S434).

  In the process of FIG. 29, the node expansion unit 17 acquires the arc 122 that exits from the node 112 (S441 to S443). The node expansion unit 17 determines that the arc 122 is not an arc to be processed (S443: NO), and the expansion processing related to the node 112 is confirmed to the simplified node 372 (“appearance frequency of co-occurrence relationship”) illustrated in FIG. To finish. This completes the processing related to “frequency”.

  Next, the selected node = node 498 (“calibration”). A node directly connected from the node 498 is a node 496 (“after”) as shown in FIG. Therefore, the node expansion unit 17 performs the process of FIG. 28 using the node 496 as a processing target node (S431). “After” is included in the noun phrase “document after proofreading” (S432: YES). At this time, the main node NA = “document” (S433). As shown in FIG. 25, the arc from “document” is arc 506 (“time”). Further, the node 496 (“after”) is expanded to “post-proofreading document” (S435).

  In the process of FIG. 29, the arc 506 (“time”) is not a process target arc (S443: NO). Therefore, the node 496 (“after”) is expanded to the minimum noun phrase “post-proofread” including the main word.

  In the above description, the node 498 (“proofreading”) is extended to the “post-proofreading document”. However, for example, in FIG. 21, the degenerate node 334 (“post-proofreading document”) is more than that. It is also possible to perform processing for determining that the data is not expanded.

  Further, as shown in FIG. 25, a node 502 (“element”) is also connected to the node 117 (“co-occurrence relationship”) connected to “the document after proofreading”. When expanding nodes, in the example in which all connected nodes are displayed as described above, noun phrases are expanded with respect to the node 502 (“element”) to obtain “each element”, and the simplified node 348 in FIG. Is displayed.

  In the input sentence 80, “in the proofread document” and “for each element” are also connected as a character string. Therefore, as shown in FIG. 23, it can also be displayed collectively as a simple node 379 (“each element in the proofread document”). As a result, a display result such as a graph 390 in FIG. 24 is obtained.

  As described above, when the user selects “frequency and proofreading”, in the present embodiment, the language processing apparatus 1 automatically displays “appearance of a co-occurrence relationship” as in the display example 382 of FIG. Select “Frequency and ... Frequency of occurrence of co-occurrence relationship”. In addition, as in a graph 390 of FIG. 24, a graph of a corresponding portion is displayed using, for example, simplified nodes 372, 379, and 392.

  As described above in detail, in the language processing apparatus 1 according to the first embodiment, the analysis unit 5 analyzes the input sentence and generates the morphological analysis information 150, the syntax analysis information 180, and the semantic analysis information 200. . The correspondence calculation unit 21 performs calculation for associating the character string with the node based on the analysis result. The display control unit 23 controls the display unit 27 to perform, for example, graph expression.

  When the user selects a character string in the input sentence or a node in the graph, the language processing device 1 acquires the selection via the input unit 3. The correspondence calculation unit 21 specifies a node corresponding to the selected character string. The noun phrase determination unit 13 and the verb phrase determination unit 15 determine whether the specified node is included in the noun phrase or the verb phrase. When the contents of the node are included in the noun phrase or verb phrase, the node extension unit 17 determines whether or not an arc not included in the phrase is generated from the main word of the phrase, and is not included in the phrase. Whether or not the arc is further expanded is determined based on whether or not the arc is the “arc” type.

  As described above, according to the language processing apparatus 1 according to the present embodiment, when displaying a sentence analysis result, a word can be expanded and displayed as a noun phrase or a verb phrase. This makes it possible to determine whether or not the character string makes sense in the semantic analysis, and to extend the unit to be displayed together to the minimum unit that makes sense in terms of assisting reading of the character string.

  Further, in the language processing device 1, it is determined whether or not to perform further expansion based on whether or not an arc from the main noun phrase or verb phrase is included in the noun phrase or verb phrase. Can do. Further, the language processing apparatus 1 can determine whether or not to perform further expansion according to the type of the arc content indicating the relationship between words as “extended” and “non-extended”. The analysis result can be displayed in units of character strings determined not to be further expanded. Therefore, it can be a unit for displaying a character string more suitable for reading support.

  At this time, in the graph expression, for example, the phrase display optimizing unit 19 appropriately changes the structure of the nodes and arcs such as grouping a plurality of nodes to be expanded. Thereby, the display regarding the dependency between words and phrases can be made into a unit more suitable for reading comprehension support. Therefore, the meaning to be displayed can be easily determined, and the display can be more easily read.

  In the language processing device 1, the display of part-of-speech information for deciphering a sentence and the display of a similar sentence, the graph expression based on a word as a sentence analysis result, and the analysis of a complex sentence that was difficult to cope with only the degenerate display in the graph expression The result can be easily displayed. Therefore, it is possible to more appropriately support reading comprehension of long sentences that could only be captured and displayed on a global basis. In addition, it is possible to visually grasp the structure of complex sentences.

  In addition, a portion that does not need to be expanded can be displayed in a reduced manner by inputting an instruction. For parts that do not need to be extended, it is also possible to display the parts without extension. Since these can be arbitrarily combined, they can be displayed in a form that matches the needs of the user. Even when the range to be selected is not appropriate from the viewpoint of semantic analysis, the selected portion can be automatically expanded and displayed in an easy-to-understand format. With such a simple display, it is possible to appropriately support reading comprehension even for users who lack linguistic expertise.

(Modification)
Hereinafter, modified examples regarding the display method and the selection method will be described. In the present modification, the redundant description of the configuration and operation of the language processing device 1 is omitted for the same parts as those in the first embodiment. 30 and 31 are diagrams showing a modification of the display method. As shown in FIG. 30, output results 524 and 526 may be displayed as examples of display of the input sentence 522. The analysis result of the sentence 522 = “Taro goes to school” may be displayed in parallel with two nodes and their arcs instead of a graph expression, as in the output results 524 and 526.

  FIG. 31 shows another example in which two nodes and their arcs are displayed in parallel, instead of a graph representation, as in FIG. As shown in FIG. 31, the display example of the input sentence 80 may be output results 534 and 536.

  32 and 33 are diagrams showing display examples when a character string is selected. As shown in FIG. 32, the display example 550 shows an input sentence 552, an initial display 556, and a post-selection display 558. In this example, an initial display 556 before selection and a display 558 after selection are shown for the character string 554 = “based” selected in the input sentence 552. These initial display 556 and post-selection display 558 may be displayed simultaneously, or the display may be switched before and after the character string is selected.

  As shown in FIG. 33, the display example 580 shows an input sentence 552, an initial display 592, and a post-selection display 594. In this example, for the character string 554 = “based” selected in the input sentence 552, an initial display 592 before the character string selection and a display 594 after the selection are shown. The initial display 592 and the post-selection display 594 may be displayed simultaneously, or the display may be switched before and after the character string is selected.

  As described above, according to the modified example, in addition to the effects of the first embodiment, it is possible to perform display that further meets the needs of the user.

  The present invention is not limited to the embodiments described above, and various configurations or embodiments can be adopted without departing from the gist of the present invention. For example, the configurations of the morphological analysis information 150, the syntax analysis information 180, the semantic analysis information 200, and the extended information 230 are not limited to the above example. Anything that shows similar contents can be applied.

  The selection of the character string in the input sentence may directly select the displayed semantic structure node. At that time, the noun phrase determination unit 13 and the verb phrase determination unit 15 perform noun phrase determination and verb phrase determination according to the part of speech of the node. These phrases may be determined based on statistical processing such as an N-gram model or other conventional methods without using the parsing information 180.

  In the embodiment and the modification described above, the case where the analysis unit 5 analyzes the input sentence has been described as an example. However, the process of performing the node expansion process based on the already analyzed result and optimizing the display May be performed. At this time, the analysis unit may acquire the morphological analysis information 150, the syntax analysis information 180, the semantic analysis information 200, and the like, which are the analysis results of the input sentence, for example, via an external storage device or a transmission / reception device. . In addition, for example, it is preferable to previously store a dictionary for performing analysis in the analysis result storage unit 25.

  In the above-described embodiment and modification, the noun phrase determination unit 13 and the verb phrase determination unit 15 are examples of phrase determination units, the node expansion unit 17 is an example of expansion units, and the phrase display optimization unit 19 is This is an example of a display processing unit. The display control unit 23 and the display unit 27 are examples of the display unit. The analysis result storage unit 25 is an example of a storage unit, and the input unit 3 is an example of a selection acquisition unit.

  Here, an example of a computer that is commonly applied to cause a computer to perform the operation of the language processing method according to the first embodiment and the modification will be described. FIG. 34 is a block diagram illustrating an example of a hardware configuration of a standard computer. As shown in FIG. 34, a computer 600 includes a central processing unit (CPU) 602, a memory 604, an input device 606, an output device 608, an external storage device 612, a medium drive device 614, a network connection device, and the like via a bus 610. It is connected.

  The CPU 602 is an arithmetic processing unit that controls the operation of the entire computer 600. The memory 604 is a storage unit for storing a program for controlling the operation of the computer 600 in advance or using it as a work area as needed when executing the program. The memory 604 is, for example, a Random Access Memory (RAM), a Read Only Memory (ROM), or the like. The input device 606 is a device that, when operated by a computer user, acquires various information input from the user associated with the operation content and sends the acquired input information to the CPU 602. Keyboard device, mouse device, etc. The output device 608 is a device that outputs a processing result by the computer 600, and includes a display device and the like. For example, the display device displays text and images according to display data sent by the CPU 602.

  The external storage device 612 is, for example, a storage device such as a hard disk, and stores various control programs executed by the CPU 602, acquired data, and the like. The medium driving device 614 is a device for writing to and reading from the portable recording medium 616. The CPU 602 can perform various control processes by reading and executing a predetermined control program recorded on the portable recording medium 616 via the recording medium driving device 614. The portable recording medium 616 is, for example, a Compact Disc (CD) -ROM, a Digital Versatile Disc (DVD), a Universal Serial Bus (USB) memory, or the like. The network connection device 618 is an interface device that manages transmission / reception of various data performed between the outside by wired or wireless. A bus 610 is a communication path for connecting the above devices and the like to exchange data.

  A program that causes a computer to execute the language processing method according to the first embodiment and its modification is stored in, for example, the external storage device 612. The CPU 602 reads a program from the external storage device 612 and causes the computer 600 to perform language processing. At this time, first, a control program for causing the CPU 602 to perform language processing is created and stored in the external storage device 612. Then, a predetermined instruction is given from the input device 606 to the CPU 602 so that the control program is read from the external storage device 612 and executed. Further, this program may be stored in the portable recording medium 616.

Regarding the above embodiment, the following additional notes are disclosed.
(Appendix 1)
Based on the analysis result obtained by extracting a plurality of words included in the input sentence and analyzing the meaning of the sentence, for each of the plurality of words, it is determined whether there is a phrase that includes the word and has one meaning. A phrase determination unit that determines and outputs the phrase if any;
As another phrase in which the word or the phrase is connected to the preceding or following word or phrase based on the relationship information indicating the relationship between two words related to each other among the plurality of words and the determination result of the phrase determination unit An extension for determining whether to display, and
When displaying the sentence, a display processing unit that performs processing to display the word or phrase determined to be displayed as the another phrase as one phrase;
Based on the analysis result and the result of processing in the display processing unit, a display unit for displaying the sentence;
A language processing apparatus comprising:
(Appendix 2)
A storage unit that stores extended information that associates the relationship information with a type classified based on whether the words or phrases should be displayed as the other phrases;
The language processing apparatus according to appendix 1, wherein the extension unit performs determination according to the extension information.
(Appendix 3)
A selection acquisition unit for acquiring that a character string in the sentence is selected;
A correspondence calculation unit that identifies the selected word or the phrase by calculating a relationship between the character string and the word or the phrase included in the character string;
Further comprising
The language processing apparatus according to appendix 1 or appendix 2, wherein the extension unit performs the determination based on the word or the phrase specified by the correspondence calculation unit.
(Appendix 4)
An analysis unit that performs morphological analysis of the sentence, performs syntax analysis of the sentence from the relationship between the morphemes based on the morphological analysis, and generates the relationship information based on a result of the syntax analysis;
Further comprising
The language processing apparatus according to any one of appendix 1 to appendix 3, wherein the extension unit performs determination based on the relationship information generated by the analysis unit as the relationship information.
(Appendix 5)
The language processing apparatus according to appendix 4, wherein the phrase determination unit classifies the word into a noun phrase or a verb phrase including the word based on a result of the parsing by the analysis unit.
(Appendix 6)
The display unit
When displaying the sentence including the word or the phrase and the relationship information, the plurality of words or the phrase that can be displayed in a reduced manner are determined and the plurality of the words that can be displayed in the reduced state or In the phrase, the plurality of words or the combination of the phrases that are simultaneously expanded to be in a non-degenerate state are determined based on the relationship information, and an instruction to switch between the reduced display and the non-degenerate display is obtained. Then, the language processing apparatus according to any one of appendix 1 to appendix 5, further comprising a display control unit that switches the display based on the instruction and the combination.
(Appendix 7)
The language processor
Based on an analysis result obtained by extracting a plurality of words included in the input sentence and analyzing the meaning of the sentence, it is determined whether or not there is a minimum unit phrase including the word and forming one meaning. And if there is, output the phrase,
Based on the relationship information indicating the relationship between two related words of the plurality of words and the determination result of the phrase determination unit, the word or the phrase is displayed as another phrase connected to the preceding or following word or phrase To determine whether or not
When displaying the sentence, performing the process of displaying the word or phrase determined to be displayed as the another phrase as one phrase,
A language processing method, wherein the sentence is displayed based on the analysis result and the processing result to be displayed.
(Appendix 8)
The appendix 7 is characterized in that the determination is performed according to extended information that associates the relationship information with a type classified based on whether the word or the phrases should be displayed as the different phrases. Language processing method.
(Appendix 9)
further,
Obtaining that the character string in the sentence is selected, and calculating the relationship between the character string and the word or the phrase included in the character string, thereby identifying the selected word or the phrase ,
The language processing method according to appendix 7 or appendix 8, wherein the discrimination is performed based on the identified word or phrase.
(Appendix 10)
Furthermore, morphological analysis of the sentence is performed,
Based on the morpheme analysis, perform a syntax analysis to analyze the structure of the sentence from the relationship between the morphemes,
Performing semantic analysis to generate the relation information based on the analysis result of the syntax analysis,
The language processing method according to any one of appendix 7 to appendix 9, wherein the determination is performed based on the relation information and the extended information generated by the semantic analysis.
(Appendix 11)
The language processing method according to appendix 10, wherein the phrase classification is performed by classifying the word into a noun phrase or a verb phrase including the word based on an analysis result by the syntax analysis.
(Appendix 12)
The processing to display is
When displaying the sentence including the word or the phrase and the relationship information, the plurality of words or the phrase that can be displayed in a reduced manner are determined and the plurality of the words that can be displayed in the reduced state or In the phrase, the plurality of words or the combination of the phrases that are simultaneously expanded to be in a non-degenerate state are determined based on the relationship information, and an instruction to switch between the reduced display and the non-degenerate display is obtained. Then, the language processing method according to any one of appendix 7 to appendix 11, wherein the display is switched based on the instruction and the combination.
(Appendix 13)
Based on an analysis result obtained by extracting a plurality of words included in the input sentence and analyzing the meaning of the sentence, it is determined whether or not there is a minimum unit phrase including the word and forming one meaning. And if there is, output the phrase,
Based on the relationship information indicating the relationship between two related words of the plurality of words and the determination result of the phrase determination unit, the word or the phrase is displayed as another phrase connected to the preceding or following word or phrase To determine whether or not
When displaying the sentence, performing the process of displaying the word or phrase determined to be displayed as the another phrase as one phrase,
A program that causes a computer to execute a process of displaying the sentence based on the analysis result and the result of the process to be displayed.
(Appendix 14)
The appendix 13 is characterized in that the determination is performed according to extended information that associates the relationship information with a type classified depending on whether the word or the phrase should be displayed as the another phrase. Program.
(Appendix 15)
further,
Obtaining that the character string in the sentence is selected, and calculating the relationship between the character string and the word or the phrase included in the character string, thereby identifying the selected word or the phrase ,
The program according to appendix 13 or appendix 14, wherein the determination is performed based on the specified word or phrase.

DESCRIPTION OF SYMBOLS 1 Language processing device 3 Input part 5 Analysis part 7 Morphological analysis part 9 Syntax analysis part 11 Semantic analysis part 13 Noun phrase determination part 15 Verb phrase determination part 17 Node expansion part 19 Phrase display optimization part 21 Corresponding calculation part 23 Display control part 23 25 Analysis result storage unit 27 Display unit 92 to 98 Degenerate node 100 to 106 Arc 150 Morphological analysis information 152 Morphological number 154 Morphological unit 156 Start position 158 Character length 160 Part of speech information 162 Meaning symbol 170 Part of speech symbol information 172 Part of speech symbol 174 Meaning 180 Parsing Information 182 Component 184 Part of speech 186 Label 188 Semantics 200 Semantic analysis information 202 Departure node 204 Arrival node 206 Arc 208 Morpheme number 210 Semantic symbol 212 Morpheme 214 Morpheme number 216 Semantic symbol 218 Morpheme 372 Simplified node 379 Simplified node 380 Display Example 382 Display example 390 Display example 392 Simplified node

Claims (6)

Based on the analysis result obtained by extracting a plurality of words included in the input sentence and analyzing the meaning of the sentence, for each of the plurality of words, it is determined whether there is a phrase that includes the word and has one meaning. A phrase determination unit that determines and outputs the phrase if any;
Perform morphological analysis of the sentence, perform syntax analysis of the sentence based on the relationship between the morphemes of the sentence based on the morphological analysis, and based on the result of the syntax analysis , semantic relationships between words, and an analysis unit for generating the relationship information indicating the semantic relation between a word that holds the primary significance in multiple clauses output from the phrase determination unit and each of the words,
Based on said relationship information, the determination of whether to display another phrase the word or the phrase concatenation and before and after the word or phrase, the relationship between the availability of it to be displayed as the another phrase said relationship information An extension unit that performs in accordance with predefined extension information ;
A display processing unit that collects the words or phrases that are determined to be displayed as the different phrases as one phrase;
Based on the analysis result obtained by analyzing the meaning of the sentence and the processing result in the display processing unit, the phrase analyzed as the central concept of the sentence , the phrase compiled by the display processing unit, and the phrase and the phrase A display unit for displaying the relationship information representing a semantic relationship with a phrase ;
A language processing apparatus comprising: A selection acquisition unit for acquiring that a character string in the sentence is selected;
A correspondence calculation unit that identifies the selected word or the phrase by calculating a relationship between the character string and the word or the phrase included in the character string;
Further comprising
The extensions, language processing apparatus according to claim 1, characterized in that the determination as an object the corresponding calculation unit the word or the phrase specified by. The language processing according to claim 1 or 2 , wherein the phrase determination unit classifies the word into a noun phrase or a verb phrase including the word based on a result of the parsing by the analysis unit. apparatus. The display unit
When an instruction to switch the display of the grouped phrases is acquired by the display processing unit, instead of displaying the grouped phrases, a word holding the main meaning in the grouped phrases, the main meaning in the grouped phrases And the relationship information representing the semantic relationship between the word or phrase other than the word holding the main word and the word or phrase other than the word holding the main meaning and the word or phrase other than the word holding the main meaning. language processing apparatus according to further comprising a display control unit for displaying on the basis of the result of the process from claim 1, wherein in any one of claims 3. The language processor
Based on the extracted plurality of words included in the input sentence analysis result of the analysis of the meaning of the sentence, for each of a plurality of said word, whether there is a phrase makes sense one includes the words Judgment, if there is, output the phrase,
Perform morphological analysis of the sentence,
Based on the morphological analysis, syntactic analysis of the sentence from the relationship between the morphemes of the sentence,
Based on the result of the parsing, a semantic relationship between two mutually related words out of the plurality of the words, and a word having a main meaning in each of the plurality of the words and the output phrase, Generate relationship information that shows the semantic relationship of
Based on said relationship information, the determination of whether to display another phrase the word or the phrase concatenation and before and after the word or phrase, the relationship between the availability of it to be displayed as the another phrase said relationship information Is performed according to predefined extended information ,
The words or phrases that are determined to be displayed as the different phrases are grouped as one phrase,
Based on the results of collectively Ru processing as the analysis result and the one phrase meaning analyzed the sentence, phrase parsed as central idea of the sentence, the gathered phrase, and, between the phrase and該句 A language processing method characterized by displaying the relationship information representing a semantic relationship . Based on the extracted plurality of words included in the input sentence analysis result of the analysis of the meaning of the sentence, for each of a plurality of said word, whether there is a phrase makes sense one includes the words Judgment, if there is, output the phrase,
Perform morphological analysis of the sentence,
Based on the morphological analysis, syntactic analysis of the sentence from the relationship between the morphemes of the sentence,
Based on the result of the parsing, a semantic relationship between two mutually related words out of the plurality of the words, and a word having a main meaning in each of the plurality of the words and the output phrase, Generate relationship information that shows the semantic relationship of
Based on said relationship information, the determination of whether to display another phrase the word or the phrase concatenation and before and after the word or phrase, the relationship between the availability of it to be displayed as the another phrase said relationship information Is performed according to predefined extended information ,
The words or phrases that are determined to be displayed as the different phrases are grouped as one phrase,
Based on the results of collectively Ru processing as the analysis result and the one phrase meaning analyzed the sentence, phrase parsed as central idea of the sentence, the gathered phrase, and, between the phrase and該句 A program for causing a computer to execute processing for displaying the relationship information representing a semantic relationship .